By Topic

On adaptive control processes

Sign In

Cookies must be enabled to login.After enabling cookies , please use refresh or reload or ctrl+f5 on the browser for the login options.

Formats Non-Member Member
$31 $13
Learn how you can qualify for the best price for this item!
Become an IEEE Member or Subscribe to
IEEE Xplore for exclusive pricing!
close button

puzzle piece

IEEE membership options for an individual and IEEE Xplore subscriptions for an organization offer the most affordable access to essential journal articles, conference papers, standards, eBooks, and eLearning courses.

Learn more about:

IEEE membership

IEEE Xplore subscriptions

2 Author(s)
Bellman, Richard ; Rand Corporation, Santa Monica, CA, USA ; Kalaba, R.

One of the most challenging areas in the field of automatic control is the design of automatic control devices that 'learn' to improve their performamce based upon experience, i.e., that can adapt themselves to circumstances as they find them. The military and commercial implications of such devices are impressive, and interest in the two main areas of research in the field of control, the USA and the USSR, runs high. Unfortunately, though, both theory and construction of adaptive controllers are in their infancy, and some time may pass before they are commonplace. Nonetheless, development at this time of adequate theories of processes of this nature is essential. The purpose of our paper is to show how the functional equation technique of a new mathematical discipline, dynamic programming, can be used in the formulation and solution of a variety of optimization problems concerning the design of adaptive devices. Although, occasionally, a solution in closed form can be obtained, in general, numerical solution via the use of high-speed digital computers is contemplated. We discuss here the closely allied problems of formulating adaptive control processes in precise mathematical terms and of presenting feasible computational algoritbms for determining numerical solutioms. To illustrate the general concepts, consider a system which is governed by the inhomogeneous Van der Pol equation \ddot{x} + \mu(x^{2} - 1) dot{x} + x = r(t), 0 \leq t \leq T , where r(t) is a random function whose statistical properties are only partially known to a feedback control device which seeks to keep the system near the unstable equilibrium state x = 0, \dot{x} = 0 . It proposes to do this by selecting the value of μ as a function of the state of the system at time t , and the time t itself. By observing the random process r(t) , the controller may, with the passage of time, infer more and more concerning the statistical properties of the function r(t) and thus may be expected to improve the quality of its control decisions. In this way the controller adapts itself to circumstances as it finds them. The process is thus an interesting example of adaptive control, and, conceivably, with some immediate ap- plications. Lastly, some areas of this general domain requiring additional research are indicated.

Published in:

Automatic Control, IRE Transactions on  (Volume:4 ,  Issue: 2 )